Hannah Frick
Hannah Frick
Following on from https://github.com/tidymodels/poissonreg/issues/82: This affects all glmnet models we wrap in tidymodels (not just here in poissonreg) where we set `relax = TRUE`. Here is an example with `linear_reg()`...
The documentation for the parsnip encodings (aka those set by `parsnip::set_encodings()`) could/should be extended to give more guidance on how to use them. When looking at what we do across...
When we talk about "model specification" and similar objects, we could/should add something along the lines of "object of type `model_spec`" to make it easier to link to the objects...
Following on from #877, the glmnet methods for `predict_numeric()`, `predict_class()`, and `predict_classprob()` would only evaluate the model spec before handing off to the method for `model_fit`. Let's check if that...
The formatting functions `format_num()`, `format_class()`, etc are fairly similar. Let's check if a general `format_parsnip()` makes sense.
This is an example for glmnet but this should probably apply to all models/engines. For glmnet, `multi_predict()` currently ignores `type = "raw"` when applied to a linear regression, but errors...
I think there is a little room to improve error messages for when what's in `new_data` does not match up with what the model was fit on. When we fit...
Originally surfaced in https://community.rstudio.com/t/error-with-tuning-for-censored-regression/181248, here is a more minimal reprex. ``` r library(censored) #> Loading required package: parsnip #> Loading required package: survival set.seed(1) lung